Used-car buyers often reveal their biggest hesitations - from loan anxieties to mechanical complaints - while behind the wheel, but those insights rarely make it back to the dealership. Indian used-car retailer Cars24 is changing that by deploying a Cars24 AI test drive system powered by ElevenLabs, turning casual in-car chatter into a massive data engine that has already driven a 35% increase in sales conversions.
Across more than 300 retail hubs, the company has analyzed over 22,000 hours of audio. According to a case study published by ElevenLabs, this surveillance-to-sales pipeline not only boosted conversions but also improved customer satisfaction by 20% and reduced hub operating costs by 10%. With 70% of Cars24 customers purchasing their first vehicle, capturing these localized, language-specific conversations has proven critical to closing deals.
How the Cars24 AI Test Drive System Works
The technical setup relies on ambient recording and automated processing to catch missed sales steps and vehicle defects.
- Ambient Recording: Tablets with open microphones are placed in the vehicle to record the entire hour-long test drive.
- Speech-to-Text: ElevenLabs software transcribes the audio and separates the speakers to distinguish the buyer from the salesperson.
- LLM Tagging: A large language model categorizes the conversation by customer, salesperson, specific car model, and retail hub.
- Managerial Review: The system flags mechanical problems noticed mid-drive and loan objections, allowing managers to compare patterns across regions.
Cars24 employs more than 1,300 contact center agents, and nearly half of the retailer's sales involve an assisted phone call. AI voice agents have now handled more than 3 million minutes of these interactions, answering buying questions and responding to objections. This automation has slashed calling costs by 50%, and the company aims to eventually route 30% of all calls through AI.
Retailers Look Past Cost Cutting
The automotive sector is not alone in pointing voice AI directly at revenue generation. Home Depot recently piloted a system in 50 stores where AI agents understand a caller's intent within 10 seconds, resolving issues four times faster than traditional phone menus.
Similarly, Japanese appliance chain Yamada Denki deployed an avatar-based voice agent that asks follow-up questions like a human associate. During a two-week campaign with 30,000 users, 92% of survey responses were positive, with shoppers noting they felt less sales pressure when discussing budgets.
Customers do not want a chatbot. They want intelligence.
- Akira Fukabori, CEO, avatarin
The Privacy Trade-Off in Physical Retail
The deployment of open microphones inside test-drive vehicles marks an aggressive expansion of retail data collection. While the 35% conversion lift and 50% drop in calling costs prove the financial viability of ElevenLabs' voice AI, it fundamentally changes the nature of a test drive from a private evaluation into a monitored sales funnel. With 43% of surveyed retailers already piloting AI shopping agents and 81% trusting them to operate autonomously, ambient surveillance is rapidly becoming a standard retail tool.
The fact that 45% of consumers are now comfortable letting an AI agent complete a purchase - rising to 54% among Gen Z - suggests buyers are increasingly willing to trade conversational privacy for frictionless transactions. As Cars24 plans to extend this technology into car inspections and financing, the line between customer support and data extraction will only blur further.
